The Research Of Trust-based Collaborative Filtering Algorithm On E-commerce Recommender System

Data continue expanding with the development of Internet informationtechnology. It becomes more and more difficult for users to find out theinterested information from great data quickly and accurately. E-commerceRecommendation System rise for recommending the interested item to users.Collaborative Filtering Recommendation technology is used the most commonly. Butit has data sparsity, cold start, scalability and other problems. These problemsreduce the quality of recommendation service.This article proposes a Trust-based Collaborative Filtering Recommendationalgorithm to solve the problems of the traditional Collaborative FilteringRecommendation algorithm. On the one hand, we add the trust relationship to thetraditional Collaborative Filtering Recommendation algorithm and it will beconsidered as a recommendation factor. On the other hand, trust can connect userswithout interaction and match more neighbors for cold start users. The Trust-basedCollaborative Filtering Recommendation algorithm integrates the similarity and trustof users fully, improves the quality of recommendation service.In this article, experiments prove that trust is viable and it can reduce datasparsity and mitigate cold start problem. The Mean Absolute Error of Trust-basedCollaborative Filtering Recommendation algorithm is smaller than the traditionalCollaborative Filtering Recommendation algorithm. The Trust-based CollaborativeFiltering Recommendation algorithm improves the recommendation service quality.